Skip to content

Experience

Tristan Jessup

Applied AI systems & integration · Senior bench technician, Geek Squad

Lexington, KY area · tristan@mindlink.dev

Most senior bench technician at Geek Squad (Best Buy Store #259), CompTIA A+, Network+ and Security+ certified and an Apple Certified Repair Technician. I design agentic tooling that pairs small language models with scripted tools, and build AI governance experiments where every claim is mapped to a test. Seeking enterprise AI integration, AI solutions and AI architecture roles.

Download resume (PDF)

Work history

Advanced Repair Agent

Apr 2021 – Present

Best Buy / Geek Squad, Store #259

  • Most senior bench technician in the store; mentor newer associates and take the most complex escalations.
  • Diagnose and repair hardware and software faults across Windows and macOS, including data recovery, NVMe/SSD cloning and migration, BIOS/UEFI issues, and malware remediation.

Founder & independent researcher

2025 – Present

Mindlink Research Group

  • Build and publish AI governance and multi-agent research code in the open, including MCX and Emergent Divergence.
  • Author of the Concord of Coexistence and the Meta-Concord draft 0.4 white paper.

Earlier roles

  • Sales · Tesla
  • City Letter Carrier · United States Postal Service
  • Customer Service Specialist · Top Drawer Furniture (furniture retail)

Selected AI work

MCX: Meta-Concord reference implementation

github.com/topstolenname/mcx-experiment

Python (Apache-2.0) reference implementation of domain-assent authorization for agent actions: Ed25519-signed evidence, a separately written verifier, a fail-closed enforcer, a red-team catalogue, and CLAIMS.md mapping each white-paper claim to a test or marking it not shown.

Emergent Divergence

github.com/topstolenname/emergent-divergence

Controlled multi-agent LLM experiment (3 conditions × 2 backends × 4 seeds) with eight falsifiable hypotheses, a decoding-noise null baseline, full run provenance, and a hard cost cap. Pilot results published, including a non-significant trend.

Anvil (paused prototype)

mindlink.dev/work/anvil

Architecture for a bench-diagnostics assistant: Markdown skill library, small language models first with escalation to larger tiers, and PowerShell tools behind a structured JSON contract.

Smaller builds

Local photo auto-tagging (Ollama vision model → exiftool metadata → Dolphin extended attributes); Home Assistant automation and a security audit of a third-party vehicle-integration repository.

Certifications

  • CompTIA A+
  • CompTIA Network+
  • CompTIA Security+
  • Apple Certified Repair Technician

Education

B.S. Information Technology, AI concentration

Strayer University

In progress

General studies coursework

University of Kentucky

2008 – 2010

Skills

AI systems

  • Agentic tool use & orchestration
  • Small and local models (Ollama)
  • Multi-agent experiment design
  • LLM evaluation & behavioural analysis
  • AI governance & authorization design
  • Claim-to-evidence documentation

Developer tools

  • Claude Code
  • Cursor
  • MCP
  • Open-WebUI
  • Python
  • PowerShell
  • Bash
  • Git
  • Docker
  • Obsidian

Systems & hardware

  • Windows & macOS diagnostics
  • Linux (Kubuntu, Arch, CachyOS)
  • Data recovery
  • NVMe/SSD cloning & migration
  • BIOS/UEFI
  • Endpoint security & malware remediation
  • Networking